Test a Small Online Product Brand Before Building a Full Store

Tom — Singapore Digital Business & Systems Consultant

Tom

Digital Business & Systems Consultant · Singapore

7 min read Article date First published
Test a Small Online Product Brand Before Building a Full Store

To test a product idea before an online store, sell one small batch through channels you already have and count every cost, not just the sales: what each unit costs once it reaches you, packaging, delivery, fees, refunds, unsold stock and your own time. Revenue is not profit; a test can bring in hundreds of dollars and still lose money. Build a full store only when the numbers work and the test has shown a problem a store would solve.

Key takeaways

  • Revenue is not profit: subtract refunds, delivery, fees, unsold stock and your time.
  • Price from landed cost: what one unit costs once it reaches you.
  • A test should show the result after all costs, effort per order, return reasons and repeat buying.
  • Choose manual orders, a marketplace or your own store by the problem you have actually hit.

Write the product and customer hypothesis

Start with a sentence you can prove wrong: "[Customer] will pay [price] for [product] because [reason], and some will buy again." Naming the price matters: people who like a product at one price may not buy it at the price your costs require.

Not sure the idea deserves a test? Start with how to test a business idea before paying for a website. Making the product yourself and taking orders first? My guide to a home pre-order round covers wording, caps and refunds. This guide covers the money and work behind each order.

Test one product in few variants; every extra colour or size leaves more stock unsold. If a supplier sets a minimum order, could you afford to lose it all? This guide uses a non-food example on purpose: food, cosmetics, children's products and anything that may need a licence have their own rules, so check the relevant authority first.

Check landed costs and fulfilment

Landed cost: what one unit really costs you

Landed cost is the supplier's price plus inbound shipping and any other charges to get the goods to you, divided by the units bought. If you make the product, count materials and labels.

Fulfilment: packaging, delivery and time

Every order needs a mailer, filler, a label and a delivery. Time your first five orders end to end, messages and drop-off included. Free delivery for buyers means that cost is yours on every order.

Returns and unsold stock

Some items will come back. Write your returns policy before the first sale, set aside an allowance for refunds and return postage, and record each reason. Stock left in the box is money spent, not profit waiting: count it as a cost until it sells.

Your unit-economics and fulfilment worksheet

Copy these lines into a spreadsheet and fill them in from real receipts. Every figure in the third column is invented, for a fictional test of 50 laptop sleeves.

LineHow to work it outFictional test (invented figures)
Landed cost per unit(Supplier price + inbound shipping and charges) ÷ units bought($500 + $100) ÷ 50 = $12
Sales before returnsOrders × price30 × $32 = $960
RefundsReturns × price; before the test, allow for about 1 order in 103 × $32 = $96
Stock in kept ordersOrders kept × landed cost27 × $12 = $324
PackagingOrders sent × cost per parcel30 × $2 = $60
Delivery(Parcels sent + return labels you pay for) × rate33 × $6 = $198
Payment or marketplace feesPlaceholder: use your provider's current published fees30 × $2 = $60
Unsold stock written offUnits left × landed cost23 × $12 = $276
ResultSales minus every cost line−$54, a loss
Your timeMinutes per order and per return, plus set-up30 × 20 min + 3 × 40 min + 12 h = 24 h, unpaid

Per order, this test looked healthy: $32 minus $12 stock, $2 packaging, $6 delivery and $2 fees leaves $10, so 30 orders look like $300. The whole test lost $54. The $354 gap is $240 for the 20 sleeves that never sold, $96 refunded on three returns and $18 of return postage. Sales of $960 and a loss can both be true; nothing pays for the 24 hours.

A fictional test: 50 laptop sleeves

(Fictional example) Jun Hao is a second-year degree student whose weekend retail shifts taught him to count stock and handle exchanges calmly. A supplier made his felt laptop sleeve in two sizes and two colours; he ordered the minimum of 50 and sold to friends, his course chat group and his social media followers over six weeks.

The upside: 30 orders, eleven from people he did not know, and two repeat buyers. The effort and uncertainty: three sleeves came back because they did not fit, the smaller grey sleeve barely sold, the worksheet showed a $54 loss, and the 24 hours fell during term. Mostly friendly buyers say little about strangers, and the 23 leftovers may never sell.

His decision: no store yet. He will list which laptop models each size fits, test a higher price and reorder only if the supplier accepts a smaller run of the size that sold.

Choose a store only when it solves a proven need

A full store is a tool for a product that has passed its test. Look for this evidence first:

  • The whole-test result is positive, or I can name the change that would make it so.
  • I know my time per order and can keep it up alongside study, work or family.
  • I know why items came back, and the main reason has a fix.
  • Some buyers came back, or strangers ordered without knowing me.
RouteSuits you whenWatch for
Manual: chat, an order form, one payment methodOrders are few and the product or price is still changingCopying details by hand; unmatched payments
A marketplace listingYou want its shoppers and can accept its rulesIts fees and rules; usually less direct contact with buyers
Your own storeProducts and prices have settled, the numbers work and buyers returnBuild and upkeep costs; finding your own visitors

For a fuller comparison, read own store vs marketplaces. Put any channel's fees into your worksheet first: more orders can still leave less on each one.

If your test has passed and manual ordering is now the obstacle, it is worth discussing e-commerce development. I build stores end to end: storefront, payment gateway setup, order tracking, customer email notifications and a backend order structure, scoped and quoted as a separate project. You keep the product, the supplier, the customers and the delivery promises. If the strain is keeping stock and orders straight, operations digitalisation may fit better.

Keep records from the first order

Keep every receipt and log each sale and refund. IRAS says income from online activities carried out on your own account for a profit is taxable, part-time or full-time, and that records should be kept for five years. If selling becomes regular, read ACRA's registration requirements: regular for-profit activity, including selling online or through social media, generally needs registration unless exempted.

Sources (checked 31 October 2026)

Frequently Asked Questions

Do I need a marketplace or my own store?

Not for a first test. Use channels you already have, then consider a marketplace if you want its shoppers and accept its rules and fees, or your own store once products, prices and repeat buyers have settled.

Does revenue mean the product test is profitable?

No. Subtract refunds, packaging, delivery, fees and every unit you bought, including unsold ones. The fictional test here took $960 and still lost $54.

How do I allow for returns before I have any?

Pick a cautious allowance, such as one order in ten, and price to cover a refund plus return postage at that rate. Afterwards, replace it with your real figure.

Should I count my own time as a cost?

Count the hours even if you do not pay yourself. A product that only works because your time is free is not yet something you can grow.

What should I do with stock that did not sell?

Treat it as a cost until it sells. Reorder only the sizes or colours that sold, and avoid a large minimum order until demand is steady.

Has a small test shown that your product pays its way, and that handling orders by hand is now the obstacle? Message me your worksheet figures and where the manual process keeps breaking. It is a conversation with no commitment, and if another test round makes more sense, I will say so.

No obligation. No sales pitch. Just an honest conversation.

Contact